The accumulation of abnormal fibrous organelles comprised of paired, helically wound intermediate filaments (PHF) represents the principal structural alteration of neuronal cell bodies and neurites during aging of the human brain and particularly in presenile and senile dementia of the Alzheimer type (Alzheimer's disease, AD). This laboratory has recently developed techniques for isolating PHF from human brain and has discovered unusual and previously unexpected molecular properties of these neuronal fibers. Our finding that PHF are highly insoluble, rigid, high molecular weight polymers has led directly to: (a) a method for preparing relatively pure PHF fractions in yields suitable for further studies; (b) the hypothesis that PHF result from abnormal crosslinking of neuronal proteins by nondisulfide covalent bonds; and (c) the production of a polyclonal antiserum which is sensitive and highly specific for PHF and fails to react with normal fibrous proteins. Based on these new findings, we now propose an integrated series of molecular studies which has as its goal the full characterization of the origin and composition of PHF. First, we shall use biochemically purified PHF as immunogen to raise monoclonal antibodies. These and the polyclonal AlphaPHF antibodies already raised will be used as specific ligands in immunoaffinity chromatography to purify PHF to homogeneity (judged by EM), a prerequisite for accurate compositional analysis of an insoluble organelle. Using purified PHF, several protein chemical strategies for structural characterization will be carried out: (a) total amino acid analyses; (b) modification of non-covalent interactions by chaotropic solvents (e.g., guanidine SCN) followed by enzymatic hydrolyses (e.g., trypsin, bacterial proteases); (c) cyanogen bromide cleavage; (e) a solid phase automated sequence analysis as well as HPLC mapping of the peptides derived from (b)-(d); (f) isolation and characterization of any acid-labile or stable crosslinks (including glu-lys dipeptide) after appropriate enzymatic or acid hydrolyses; (g) analysis of non-protein constituents of PHF. Monoclonals to PHF will also be employed in immunohistochemistry (light and EM) and immunoblotting to delineate further the distribution of PHF in AD tissue and within neurons and their processes and to identify neuronal and non-neuronal proteins with shared determinants. These various studies should provide new information about the origin of PHF and the preceding abnormal biochemical events in Alzheimer neurons that allow their formation.